FURI | Spring 2026

Memory-Inspired Modeling of Alzheimer’s Progression: A KG – Driven Approach

Data icon, disabled. Four grey bars arranged like a vertical bar chart.

This research aims to overcome the temporal blindness of standard artificial intelligence (AI) models by developing an architecture equipped with longitudinal memory to accurately forecast Alzheimer’s disease progression. By synthesizing timelines from 1,730 Alzheimer’s Disease Neuroimaging Initiative (ADNI) patients into a novel Patient-Trajectory Knowledge Graph, the memory-enabled model successfully tracked cognitive decline, whereas stateless baselines failed. This provides society with a reliable prognostic tool to anticipate trajectories and tailor early clinical interventions. Future iterations will transition this architecture into an autonomous diagnostic tool capable of immediate, data-backed prognostic forecasting for clinicians.